Key Influencing Factors in the Variation in Livestock Carbon Emissions in the Grassland Region of Gannan Prefecture, China (2009–2024)

Abstract

Research was conducted in Gannan Prefecture, China, to better understand the characteristics of carbon emissions and sequestration in areas dominated by animal husbandry. The emission factor method was used to calculate and analyze changes in carbon emissions from 2009 to 2024. The region’s average annual carbon emissions from animal husbandry are 774,286 t C-eq (2,839,049 t CO₂eq), with enteric emissions from cattle being the biggest contributor. However, as the number of locally raised cattle and sheep has decreased, carbon emissions have gradually fallen at an average annual rate of −1.0%. The annual average total carbon sequestration of vegetation in the region is 6,861,535 t C-eq, and the carbon content in underground biomass is higher than that in aboveground biomass, making it the main contributor to grassland carbon sequestration. Carbon sequestration from grassland vegetation is greater than the carbon emissions from animal husbandry, which means that the entire production system is currently a carbon sink. Meanwhile, the analysis of land-use carbon sequestration found that the annual average total sequestration by forests and grasslands over the same time period was 752,327 t C-eq, and sequestration is increasing at an annual rate of 1.4%, primarily driven by the progressive expansion of forested areas. Although the regional carbon emissions from animal husbandry are lower than the carbon sequestration, developing a science-based animal husbandry plan aligned with regional ecological thresholds, continuing to implement grass–livestock balance management measures, and preventing livestock numbers from exceeding their ecological carrying capacity remain critical to promoting sustainable coordination between livestock economies and ecological conservation.

1. Introduction

Climate change, driven primarily by human activities, is becoming increasingly severe at the global scale and has emerged as a pressing global challenge [1,2]. An assessment of the relationship between global climate change and greenhouse gas (GHG) emissions in 246 countries reveals a robust positive association between climate and GHG emissions [1]. Total GHG emissions (excluding land use, land-use change, and forestry emissions) amounted to 24,002.75 Mt CO₂eq (carbon dioxide equivalent) in 1970 and 52,962.9 Mt CO₂eq in 2023, accounting for an average annual growth rate of 1.5% during that period [3]. According to FAO statistics, total carbon emissions from agrifood systems increased from 14,208,933.8186 kt CO₂eq in 1990 to 16,240,231.6213 kt CO₂eq in 2022, with an annual average of 15,371,649.8 kt CO₂eq and an average annual growth rate of 0.4% [4]. At the same time, global emissions from livestock increased from 3,594,421.47 kt CO₂eq to 4,259,877.76 kt CO₂eq, with an annual average of 3,798,342.0 kt CO₂eq and an average annual growth rate of 0.5% [4]. During this period, livestock carbon emissions globally accounted for 24.80% of agrifood systems carbon emissions and displayed an average annual growth rate of 0.1% [4]. Globally, the livestock sector supports about 1.3 billion producers and retailers and contributes 40–50% of agricultural gross domestic product (GDP) [5]. Over 20 billion livestock animals occupy 30% of the Earth’s terrestrial surface, consuming 1/3 of cropland for feed production and 32% of freshwater resources while sustaining the livelihoods of 1.3 billion people. In the past 40 years, driven by population growth, income increase, and urbanization, the global per capita consumption of livestock products has more than doubled. To meet this demand, the production of beef, milk, and monogastric animals (pigs and poultry) has significantly increased. Total emissions from livestock between 1995 and 2005 were between 5.6 and 7.5 Gt CO₂eq yr⁻¹, and cattle production systems dominated the sector’s emissions [5]. Projections show sustained demand growth over the next two decades, posing a significant challenge to global GHG reduction efforts [5]. The carbon emissions generated by global animal husbandry cannot be ignored. Reducing GHG emissions, mitigating climate change risks, and maintaining the stability of the Earth’s ecosystems are of great significance for the sustainable development of human economy and society [6].

Emissions from China’s livestock (352,283.8 kt CO₂eq) ranked third globally between 1990 and 2022, behind India (465,310.5 kt CO₂eq) and Brazil (420,919.0 kt CO₂eq) [4]. China is a major producer and consumer of ruminant products and, from 1990 to 2022, contributed approximately 9.3% of total global livestock carbon emissions (measured as CO₂eq) from its livestock sector [4,7,8], and ruminant production generates 39% of China’s total agricultural GHG emissions [8,9]. With the development of China’s economy and rapid urbanization, the demand for livestock products is expected to increase, and sustainable animal husbandry will become a key factor in reducing GHG emissions and mitigating climate change in China’s agricultural sector [10,11]. With economic growth and rapid urbanization, the dietary pattern of Chinese residents is transforming from plant-based to animal-based consumption [12]. Empirical research demonstrates that GHG emissions associated with animal-derived food production are twice those of plant-based alternatives. This rapid expansion in animal protein consumption will undeniably contribute to a surge in GHG emissions from livestock production, posing a formidable challenge to China’s commitments to carbon reduction [13].

Studying carbon reduction in different areas of animal husbandry production is of positive significance if China is to achieve carbon neutrality and sustainable agricultural development. Existing research primarily focuses on measurements at the national level, with a notable deficiency in the understanding of the distinct carbon emission characteristics of livestock farming at the provincial level in China [13,14,15,16]. Due to substantial regional variations in natural conditions, agricultural infrastructure, and crop–livestock systems across China [13,14,17,18,19], the carbon emissions profiles of agricultural production systems exhibit distinct spatial heterogeneity. Conducting integrated analyses of region-specific planting and breeding practices is crucial to formulating spatially differentiated carbon reduction strategies. Northwest China’s livestock sector significantly contributes to carbon emissions within the agricultural production system [13,14,17]. Therefore, conducting carbon emission research that considers the specific characteristics of agricultural production in specific regions not only is more scientific and reasonable but also helps to formulate carbon reduction policies suited to local conditions.

In this study, we conducted a detailed carbon emission and sequestration analysis of Gannan Autonomous Prefecture, a typical livestock production area in Gansu Province and in northwest China, and analyzed the carbon balance of the livestock production system in the context of land use. By leveraging the region’s existing land-use patterns, we estimated the appropriate livestock inventory under carbon balance conditions. The objective is to achieve neutrality between land absorption and animal emissions, preserve the grassland ecosystem, and foster sustainable economic and social development in grassland regions, thus providing a reference for reducing agricultural carbon emissions in China’s livestock regions.

2. Materials and Methods

2.1. Study Area

Gannan Prefecture is located in the southwest of Gansu Province, in the transition zone between the northeast edge of the Qinghai Tibet Plateau and the western part of the Loess Plateau. The region has an altitude of 1100–4900 m, with most areas lying above 3000 m. The average temperature is 1.7 °C, with a short frost-free period and extensive sunshine hours, creating a typical continental climate. The region is home to vast grasslands, covering 56.13% of the total area (3.8312 million ha), and forests, which cover an additional 28.66%. Shrubland and farmland occupy smaller proportions of this region, accounting for 2.22% and 2.39% of the total area of Gannan Prefecture, respectively [20]. Economic production is dominated by grassland animal husbandry, which is the primary source of income for local residents. The output value of animal husbandry accounts for 89.14% of the output value of the primary industry. The region’s abundant natural grassland resources are realized mainly through natural and rotational grazing [20,21].

2.2. Data and Methods

2.2.1. Carbon Emission Calculation

The main livestock raised in the region are cattle, sheep, and pigs [21]. The proportion of cattle to the local large animal population fluctuated between 94 and 96% during the study period, and cattle accounted for approximately 95% of the region’s large livestock, with horses, donkeys, and mules making up the rest (5%) [21]. Based on local agricultural and animal husbandry production characteristics and corresponding data from the city’s statistical yearbook, these three species were chosen to calculate carbon emissions [21]. GHG emissions released from animal husbandry include methane (CH₄), produced by animal enteric fermentation, as well as CH₄ and N₂O emissions, generated during livestock manure management. The GHG emissions in this sector are estimated following the Intergovernmental Panel on Climate Change (IPCC) proposed methodology [22]; the calculation methods and emission factors for livestock GHG emissions are detailed elsewhere [14,19]. To compare the carbon emissions and carbon sequestration of animal husbandry, the GHGs of animal husbandry were calculated in terms of C-eq. We calculated carbon equivalence with global warming potential (GWP₁₀₀) values of CH₄ = 27 and N₂O = 273, which were obtained from Climate Change 2022, IPCC [23]. The greenhouse effect caused by 1 ton of N₂O is equivalent to the greenhouse effect caused by 74.455 tons of C, and the greenhouse effect caused by 1 ton of CH₄ is equivalent to the greenhouse effect caused by 7.364 tons of C [19].

Animal husbandry carbon emission E_Total (C-eq) is calculated as

E_CH4 = ∑(E_{CH4,enteric,i} +E_CH4,manure,i) = ∑(F_{CH4,enteric,i} ×Pi + F_CH4,manure,i × Pi)

(1)

E_N2O = F_N2O,manure,i × Pi

(2)

E_Total = E_CH4 × 7.364 + E_N2O × 74.455

(3)

where ECH₄ denoted the total CH₄ emissions from enteric fermentation and manure management (kg a⁻¹) and FCH₄ is the emission factor of CH₄ from the i-th type of animal (kg head⁻¹ a⁻¹). For enteric fermentation, this is the mean value of large-scale farming, free-range farming, and grazing farming, and for manure management, it is the calculation according to emission factors for the northwest region of China [19]. P_i (head a⁻¹) is the annual breeding amount of the i-th type of animal. EN₂O is the total N₂O emission from animal manure management (kg a⁻¹). FN₂O is the manure management N₂O emission factor for the i-th type of animal, in units of kg head⁻¹ a⁻¹. The corresponding emission factors of animals were taken from the Guidelines for the Preparation of Provincial Greenhouse Gas Inventories [14,19]. E_Total is the total GHG emission from all animals, in C-eq, t·a⁻¹ [14,19].

2.2.2. Animal Husbandry Carbon Sequestration

Grassland is an important component of the livestock production system. The carbon sink function of grassland can effectively reduce the pressure of GHG emissions arising from animal husbandry. With reference to the available literature, we refer to grassland carbon sequestration as carbon sequestration for animal husbandry development [14]. The region of Gannan is part of the Qinghai Tibet Plateau, with alpine grassland suited to low temperatures and a humid environment. Alpine grasslands have vast soil organic carbon density and storage capacity, with a carbon content of underground biomass 6.47 times that of aboveground biomass [24]. We used grassland biomass and the biomass carbon conversion coefficient to calculate grassland carbon uptake according to the following formula [14]:

W_grass (t) = (N + N× α) × β × δ

(4)

where W_grass is the carbon uptake of grassland (in C-eq, t), N is the aboveground biomass of grassland, and α is the proportional coefficient of underground and aboveground biomass. The conversion coefficient value is 6.47 [24]; β is the conversion coefficient of biomass and dry matter (0.5); δ is the carbon conversion coefficient of dry matter, using the international common conversion rate of 0.45 [14].

2.2.3. Calculation of Other Carbon Indicators

Other carbon indicators for agricultural ecosystems are measured according to carbon emission intensity (CEI), carbon sequestration intensity (CSI), and net carbon intensity (NCI), which were calculated as follows:

CEI (t·ha⁻¹) = E_Total/S_g

(5)

CSI (t·ha⁻¹) = W_grass/S_g

(6)

Net C change (t) = W_grass − E_Total

(7)

NCI (t·ha⁻¹) = Net C change/S_g

(8)

I_S = Net C change/W_grass

(9)

where S_g is the grassland area (ha) and I_s (%) is the index of sustainability [14,19].

2.2.4. Carbon Emissions from Different Land-Use Types

The accounting of land-use carbon emissions includes both direct and indirect land-use emissions. This study focuses on direct carbon emissions stemming from land-use activities. Forest land is used primarily for ecological protection, and to alleviate grassland pressure, some forest land is used to supplement animal husbandry [25]. The carbon emissions from local forests and grasslands were calculated as follows [26]:

C_s = C_g + C_f = ΣA_j μ_j

(10)

where C_s is the total direct carbon emission from land use (in C-eq, t·a⁻¹), and C_g and C_f are the carbon emissions from grasslands and forests, respectively. Total carbon emissions were derived by separately calculating emissions for these two land-use types by using the carbon emission coefficient method, followed by the summation of the individual results. A_j is the area of different types of land (hm²); j refers to grassland or forest; μ_j is the carbon emission coefficient (a negative number means carbon absorption) for each land-use type j. The carbon emissions for forest use and grassland use are −0.0644 and −0.0021 kg/(m²·a) [26].

3. Results

3.1. Changes in Livestock Carbon Emissions

From 2009 to 2024, among the three main types of animals raised in the region, namely, cattle, sheep, and pigs, the number of sheep raised decreased the most significantly, gradually dropping from 2.4343 million to 1.2992 million, representing an average annual decline rate of 4.1%. The overall number of cattle raised slowly declined from 1.266 million to 1.2064 million. During the period of 2018–2020, the number of cattle raised was slightly higher than that in other years, with an average annual decrease rate of 0.3%. The number of pigs raised was relatively stable, with an average annual growth rate of 0.7% (Figure S1).

The total carbon emissions corresponding to the CH₄ produced by enteric fermentation in all animals show a slow downward trend from 2009 to 2024, with slightly higher emissions from 2018 to 2020 compared with other periods. The annual average emission is 685,279 t C-eq, with an average annual growth rate of −1.0% (Figure 1).

An analysis of the characteristics of carbon emissions from different animals’ enteric fermentation reveals that emissions from cattle have the same trend as total emissions from all livestock in the entire region. In 2018, the maximum emissions were 622,760 t C-eq, while 2021 had the lowest emissions, at 519,188 t C-eq. The annual average emissions were 561,526 t C-eq, with an average annual growth rate of −0.3%.

Enteric fermentation emissions from sheep also show a decreasing trend year by year, with an annual average of 122,060 t C-eq and an average annual growth rate of −4.2%. Meanwhile, emissions from pigs have slowly increased, with an annual average of 1693.5 t C-eq and an average annual growth rate of 0.7%.

An analysis of the proportion of enteric emissions from each animal shows that the proportion of emissions from cattle varied between 78% and 87%, increasing year by year, with an average annual proportion of 82.1%. The proportion of emissions from sheep varied between 21% and 13%, decreasing year by year, with an average annual proportion of 17.7%. The proportion of pig emissions was below 0.3% (Figure S2).

Changes in GHG emissions corresponding to the CH₄ and N₂O generated from all animal manure management over time reveal a similar overall emission trend to that of total enteric emissions. Between 2009 and 2024, annual average fecal emissions were 89,015 t C-eq, with an average annual growth rate of 0.6% (Figure 2). An analysis of the emission characteristics of different animals reveals that the trend of emissions from cattle manure management was similar to the total emissions of all livestock. In 2018, the maximum value reached 75,995 t C-eq, and, in 2021, it was the lowest, at 63,356 t C-eq, with an average annual growth rate of −0.3%. Emissions from pig manure slowly increased year by year, with an average annual growth rate of 0.7%. Emissions from sheep generally decreased, with an average annual growth rate of −4.1%. The analysis of the proportion of emissions from each animal shows that the proportion of emissions from cattle varied between 75% and 81%, increasing year by year, with an average annual proportion of 77.1%. The proportion of emissions from sheep decreased year by year over the years, from 20% to 12%, with an average annual proportion of 16.5%. The proportion of pig emissions slowly increased year by year, with an average annual value of 6.4% (Figure S3). From 2009 to 2024, the ratio of total carbon emissions from enteric carbon to total emissions from manure remained relatively constant for all animals, with a ratio of 7.7:1.0.

Figure 3 and Figure 4 show the total amount of GHG and composition ratios from different livestock animals. According to Figure 3, total GHG emissions from all livestock decreased slowly from 2009 to 2024, with emissions from 2018 to 2020 being slightly higher than in other years. The average annual GHG emissions were 774,286 t C-eq, with an average annual growth rate of −1.0%. GHG emissions from cattle were similar to the overall regional trend, with average emissions of 630,048 t C-eq. The annual average emission from sheep was 136,869 t C-eq, with an average annual growth rate of −4.1%. The trend of GHG emissions from pigs remained stable, with an annual average of 7369.6 t C-eq and an average annual growth rate of 0.7%. Total GHG emissions from cattle in the region, calculated as carbon, were, on annual average, 4.6 times those of sheep and 85.5 times those of pigs.

The annual proportion of GHG emissions from cattle varied between 78% and 86%, showing a slow gradual increase, with an average annual proportion of 81.5%. The annual proportion of GHG emissions from pigs varied between 0.8% and 1.1% and remained relatively stable, with an average annual value of 1.0%. The proportion of GHG emissions from sheep varied between 21% and −12% and gradually decreased, with an annual average of 17.5% (Figure 4). The GHG emissions from all animals are calculated as carbon emissions. Between 2009 and 2024, the total carbon emissions from Gannan show a slow downward trend, with an average annual growth rate of −1.0%. Cattle accounted for 81.5% of carbon emissions in the region, sheep for 17.5%, and pigs for 1.0%.

3.2. Carbon Sequestration of Production System

Between 2009 and 2024, the total carbon locked in grassland biomass in Gannan showed a decreasing trend, with an annual average of 6,861,535 t C-eq and an average annual growth rate of −1.8%. Quantitative changes in carbon sequestration above and below ground were similar to the total carbon sequestration in grasslands, with annual average values of 918,546 and 5,942,990 t C-eq, respectively. Underground biomass has a higher carbon content than aboveground biomass, making it the primary contributor to grassland carbon sequestration. The carbon sequestration of aboveground biomass is equivalent in magnitude to the total emissions of all livestock animals, with an annual average of 1.2 times animal emissions. The annual average carbon sequestration of underground biomass is 7.68 times the annual average of total animal carbon emissions. Therefore, the overall carbon balance of the entire production system has greater carbon absorption than carbon emissions. The annual average carbon sequestration (i.e., net C) is 6,087,249 t, with an average annual growth rate of −1.8%, meaning that the region is a carbon sink (Figure S4).

3.3. Carbon Indicators of Production System

The total carbon emission calculation reveals that the ratio of total carbon emissions from all livestock animals in the region to the grassland area, that is, the CEI, varied from 0.32 t·ha⁻¹ to 0.47 t·ha⁻¹ from 2009 to 2024. Over time, there has been an overall slow increase trend, with the highest CEI value reaching 0.47 t·ha⁻¹ in 2020. The average annual value is 0.36 t·ha⁻¹, with an average annual growth rate of 0.9%. The average annual value of grassland CSI was 3.2 t·ha⁻¹. The NCI, which refers to the ratio of regional net carbon to grassland area, fluctuated between 2.72 t·ha⁻¹ and 2.87 t·ha⁻¹, showing an overall downward trend, with an average annual value of 2.83 t·ha⁻¹ and an average annual growth rate of −0.1%. Is ranged from 89% to 85%, with an average annual value of 89%, showing an overall downward trend and an average annual growth rate of −0.1%.

3.4. Characteristics of Land-Use Carbon Emissions

The main land-use types in this region are grassland and forest land. Therefore, we chose these two land types to calculate land-use-related carbon absorption. The results are shown in Figure 5. The average annual land carbon absorption in the grassland area was 45,163 t C-eq. As the grassland area decreased, its absorption of carbon gradually decreased, with an annual growth rate of −3.0%. The average annual carbon absorption of forest land was 707,165 t C-eq. As forest land area increased, its absorption of carbon gradually increased, with an annual growth rate of 1.7%. The area of grassland in Gannan is about twice that of forest land, but the ability of grassland to absorb carbon is much lower. The total amount of carbon absorbed by forest and grassland shows an increasing trend year by year, with an annual growth rate of 1.4% and an average annual value of 752,327 t C-eq. Before 2021, carbon emissions from livestock animals in the region were greater than the total carbon absorbed by forests and grassland. From 2022 onwards, carbon emissions from animal husbandry in the region were lower than the quantity of carbon absorbed by the land. Forest land is the main carbon sink in the local area and plays a significant role in reducing carbon emissions from animal husbandry. Since 2021, from the perspective of land use, the livestock industry in the region has achieved carbon balance.

4. Discussion

Between 2009 and 2024, the annual average carbon emissions from all animals in the local livestock production system of Gannan amounted to 774,286 t C-eq, with the proportion of carbon emissions from cattle accounting for 81.5%, those from sheep for 17.5%, and those from pigs for 1.0%. Enteric fermentation processes represented the main contributing factor to carbon emissions; in particular, the enteric emissions from cattle accounted for 73% of all animal carbon emissions. This is similar to the carbon emission characteristics of the livestock production system of Zhangye City, Gansu Province, where research found an annual mean emission average of 541,300 t C-eq from 2010 to 2021, of which enteric fermentation was the largest contributor (86%) [19]. Similarly, research in the Qinghai–Tibet Plateau from 1990 to 2015 found that cattle were also the main contributors to animal husbandry GHG emissions [13], while a study in Inner Mongolia found that from 2009 to 2022, cattle accounted for 50% of total carbon emissions [27].

A study of non-CO₂ GHG emission intensity in China’s livestock sector across 31 provinces from 2006 to 2022 revealed that enteric CH₄ emissions constituted 58.47% of total GHG output [11]. Among livestock species, cattle, pigs, and sheep/goats contributed the largest proportions of CO₂ equivalents, 44.53%, 24.07%, and 20.98%, respectively [11]. Meanwhile, a study on livestock GHG emissions and mitigation potential in China between 2000 and 2020 showed that CH₄ emissions from enteric fermentation were the main source of GHG, with these emissions generally showing a downward trend [13]. The study also found that east China had the highest GHG emissions from the livestock sector, with average annual emissions of 111.70 Mt CO₂eq, while south China had the lowest emissions, 33.70 Mt CO₂eq. In 2020, the GHG emission intensities (per unit of GDP) increased gradually from the southeast to the northwest, with Tibet, Qinghai, and Gansu having the highest emission intensities [13]. Xu et al. calculated the median total CH₄ emitted from the livestock sector in China in 2014 to be 10.8 Tg CH₄·yr⁻¹, of which the largest CH₄ emission source was beef cattle at 2.8 Tg CH₄·yr⁻¹ [28]. From 2006 to 2020, GHG emissions from livestock production in China have maintained a steady decline at approximately 170 Mt CO₂eq [29].

The annual average carbon sequestration of aboveground biomass in the grassland area of Gannan is equivalent to the annual average total carbon emissions of existing livestock, whereas the annual average carbon sequestration of underground biomass exceeds the annual average of animal carbon emissions. The carbon sequestration of underground biomass is essential to reducing carbon emissions from animal husbandry and maintain a positive carbon balance. Therefore, the region should continue to strengthen the protection of local grassland vegetation, especially in ensuring the stability of underground biomass.

The carbon emissions from animal husbandry in the region have been decreasing continuously (Figure 5), and the carbon sink level has been improving consistently. These results are closely related to the local government’s continuous promotion of grassland ecological environmental protection over the past 30 years. In the 1990s, permafrost degradation and overgrazing led to local desertification and a reduction in wetland area [30]. The local government initiated the management of grassland ecology. In 1991, the first “Grassland Management Measures” were promulgated, which clarified a system of grassland contracting responsibility, clarified the ownership of use, and prohibited sale or the change in use. At the time, overgrazing was initially curbed through the confirmation of grassland property rights; however, due to a lack of technical funding, the degraded area continued to grow, and desertification and rodent infestations were not effectively addressed [31]. In 2002, the “Grassland Management Measures” were revised to introduce a system that balanced grass and livestock, and special projects, such as returning grazing land to grass and preventing rodent infestations, were implemented [32]. The “Management Measures for Grass and Livestock Balance” were implemented in 2011 [33].

Key projects have included “using grass to determine livestock,” pilot projects for the ecological restoration of degraded grasslands through artificial grass planting, the implementation of natural vegetation protection, returning farmland to forests (grasslands), and ecological restoration and management projects. With the release of the “Implementation Plan for Grassland Ecological Protection Subsidy and Reward Policy in Gannan Prefecture” [34], the national grassland ecological subsidy mechanism was implemented, and subsidies were issued to areas where grazing had been prohibited and the grass–livestock balance was maintained. The 2023 version of the “Grassland Management Measures” further strengthens legislation prohibiting grazing and resting grazing and implements ecological governance measures such as grass and livestock balance [32]. The continued implementation of these policies has promoted a significant reduction in the livestock inventory, increased the forest area, and supported the ecological restoration of grasslands in the region. From 2009 to 2024, the inventory of cattle and sheep in the region has been decreasing year by year, with average annual growth rates of −0.3% and −4.1% in carbon emissions (Figure 3) and an average annual growth rate of 1.3% in forest area (Figure 5). In 2013, the comprehensive vegetation cover of grasslands in Gannan Prefecture was 96.3%, and by 2023, this was 97.14% for the entire state, with a cumulative potential desertification control area of 19,654 hectares [35].

The livestock industry in the region has reached a carbon-neutral state; however, according to the Calculation of Reasonable Carrying Capacity of Natural Grasslands [36], based on grassland area and aboveground grass production until 2022, the carrying capacity of large livestock in the region had always been overloaded [34,37,38]. In 2010, the theoretical carrying capacity of the region was 4.08587 million sheep units, and the actual carrying capacity was 5.45062 million livestock units, with an overload rate of 33.4% [34]. Throughout the past decade, the comprehensive vegetation coverage of the grasslands in Gannan Prefecture has been increasing year by year. In 2013, the average fresh grass yield per mu of grassland in the whole state was 371.1 kg, with a 15% increase in average fresh grass yield per mu. As a result, the theoretical carrying capacity of grasslands has continuously increased [37]. By strictly implementing the principle of “using grass to determine livestock,” since 2023, the region has achieved a dynamic balance between grass and livestock [35]. The continuous implementation and execution of these policies and measures to protect grasslands have played an important role in increasing regional carbon sinks, reducing carbon emissions from animal husbandry, achieving carbon neutrality, and coordinating the synchronous development of animal husbandry economy and ecological protection.

To mitigate global carbon emissions and curb the rise in global temperatures, in 2020, the Chinese government unveiled an ambitious goal to “attain carbon peak status by 2030 and … carbon neutrality by 2060” [39]. The varying climatic conditions, forage resources, feeding structures, and technologies, as well as the diverse supporting policies for livestock production across different regions, have contributed to distinct GHG emission characteristics [13]. Therefore, each region of China should develop scientific and reasonable carbon reduction policies based on the characteristics of carbon emissions from agricultural production in their respective regions. By continuously adhering to their implementation, these dual carbon goals can be achieved.

From the perspective of land use, scientifically and rationally utilizing land is another effective measure to reduce regional carbon emissions [40]. The carbon sink of forest land is higher than that of grassland, which is an important factor contributing to regional carbon balance. A systematic evaluation of the carbon sequestration potential of agroforestry systems for achieving agricultural carbon neutrality in Europe revealed that 64 agroforestry configurations exhibit carbon sequestration capacities ranging from 0.09 t C·ha⁻¹·a⁻¹ to 7.29 t C·ha⁻¹·a⁻¹, encompassing diverse land-use practices, from field boundary hedges to fast-growing shrub systems [41]. Implementation in priority areas could achieve carbon sequestration of 2.1 million t C·ha⁻¹·a⁻¹ to 63.9 million t C·ha⁻¹·a⁻¹ (7.78 million t CO₂eq a⁻¹ and 234.85 million t CO₂eq a⁻¹, respectively), offsetting 1.4–43.4% of Europe’s agricultural GHG emissions. In the European context, this demonstrates that optimized agroforestry configurations could significantly enhance the negative emission capacity of agriculture, providing critical technical pathways for achieving climate neutrality under Europe’s Farm to Fork Strategy.

An investigation in north China on the impact of supporting land on GHG emission intensity in coupling crop–livestock production (CLP; i.e., crops grown on supporting land to provide feed for livestock and return manure to supporting land) farms found that CLP farms with supporting land exhibit higher GHG emission intensity and greater carbon sequestration potential than single-enterprise farms (i.e., either feed is produced on supporting land or manure is returned to the land, but not both) [42]. To promote low-carbon development, the study suggests optimizing the proportion of animal land occupation, maintaining moderate animal density (e.g., approximately 144 pigs per hectare), increasing farm feed production (at least 30% self-sufficiency), and increasing the return of manure to land (at least 15 tons/hectare). This approach demonstrates that implementing a balance between grass and livestock, controlling grazing density, and implementing rotational grazing practices, as well as returning manure to grasslands, can promote the sustainable development of low-carbon animal husbandry in Gannan.

In addition, by improving feed utilization efficiency, improving feed management, reducing carbon emissions in manure management, and utilizing advanced collection and processing technologies, manure can be returned to grasslands while optimizing animal populations and alleviating environmental pressures [5]. Building low-carbon, efficient, and green livestock farms is an essential means to achieve sustainable development within these regional economies.

5. Prediction

In livestock production systems, reasonably regulating the industrial structure and optimizing the number of animals can also serve as an important means to reduce carbon emissions. Given the uncertainty regarding the future stock levels of cattle, sheep, pigs, and other livestock, extensive simulation calculations can be performed to identify the optimal stock levels of these animals, thereby achieving the goal of reducing carbon emissions year by year. To this end, this research study conducted a basic Monte Carlo simulation using key parameters within defined ranges. The calculation and simulation process are as follows.

Assuming that from the year 2025 onwards, the annual growth rate of different livestock species is α_i, the quantity p_i(t) of each livestock species in some year (t) can be calculated according to Formula (11):

$$\begin{array}{l}{p}_{\mathrm{i}}\left(t\right)=p_{\mathrm{i}}\left(2024\right){\left(1+{\alpha }_{\mathrm{i}}\right)}^{t-2024} ,\\ \mathrm{and} {\alpha }_i\in \left[a_{\mathrm{i}},b_{\mathrm{i}}\right]\left(\mathrm{i}=\mathrm{1,2},\dots ,n\right),2025\le t\le T_0\end{array}$$

(11)

where [a_i, b_i] represents the confidence interval for the average annual growth rate of the i type of livestock from 2025 onwards. a_i is set as the minimum average annual growth rate between 2010 and 2024, while b_i is the maximum during the same period, and they adhere to conditions (12) and (13), respectively:

$$a_{\mathrm{i}}=\mathrm{M}\mathrm{i}\mathrm{n}\left\{{\alpha }_{\mathrm{i}}\left(t\right)|2010\le t\le 2024\right\}$$

(12)

$$b_{\mathrm{i}}=\mathrm{M}\mathrm{a}\mathrm{x}\left\{{\alpha }_{\mathrm{i}}\left(t\right)|2010\le t\le 2024\right\}$$

(13)

where α_i (t) denotes the average annual growth rate of the i-th type of livestock in year t, with i = 1, 2, …, n, and n is the total number of livestock types. Based on historical data of different livestock, the confidence intervals for a_i and b_i of cattle, sheep, and pigs are calculated to be within the ranges of [−0.153, 0.177], [−0.353, 0.050], and [−0.037, 0.085], respectively. The core of the simulation lies in selecting α_i (i = 1, 2, …, n) to ensure a year-by-year decrease in carbon emissions from the livestock production system after 2025, as well as in identifying the relationships among the stock levels of cattle, sheep, and pigs. Taking cattle, sheep, and pigs as samples, by randomly selecting α_i ∈ [a_i, b_i] (i = 1, 2, …, n), after the p_i(t) data have been calculated, discriminate analysis is conducted by using Formulas (1), (2), and (3) from Section 2.2.1. Numerous sampling simulation calculations and data analysis indicate that the simulation results of the inventory levels of cattle, livestock, and pigs follow assumptions 1 and 2 below.

1. Assuming that from 2024 onwards, the stock level of cattle p₁(t) decreases year by year, while the stock levels of sheep p₂(t) and pigs p₃(t) increase year by year, when the stock levels of these three animals satisfy the relationship expressed in Formula (14), it can be ensured that the total annual carbon emissions from livestock farming in the region will decrease year by year from 2024 onwards.

$${\displaystyle \frac{p_2(t+1)-p_2\left(t\right)}{6}}+{\displaystyle \frac{p_3(t+1)-p_3\left(t\right)}{14}}\le p_1\left(t\right)-p_1\left(t+1\right), t>2024$$

(14)

2. Assuming that from 2024 onwards, the stock level of cattle p₁(t) increases year by year, while the stock levels of sheep and pigs decrease year by year, when the stock levels of these three animals satisfy the relationship expressed in Formula (15), it can be ensured that the annual carbon emissions will decrease year by year from 2024 onwards.

$${\displaystyle \frac{p_2\left(t\right)-p_2(t+1)}{8}}+{\displaystyle \frac{p_3\left(t\right)-p_3(t+1)}{16}}>p_1\left(t+1\right)-p_1\left(t\right), t>2024$$

(15)

where p₁, p₂, and p₃ represent the stock levels of cattle, sheep, and pigs respectively.

According to assumption 1, when the growth rates of cattle are −1%, −2%, and −3%, the growth rates of sheep are +3%, +4%, and 5%, respectively, and the fixed growth rate of pigs is +8.5%. The trend of changes in carbon emissions is depicted in Figure 6.

As shown in Figure 6, the carbon emissions of cattle have the greatest impact on the trend of total carbon emissions. The results indicate that in 2030, when the annual growth rates of cattle, sheep, and pigs are −1%, +5%, and +8.5%, respectively, compared with 2024, the total carbon emissions will begin to rise. To ensure that carbon emissions do not increase after 2024, the annual growth rates of cattle, sheep, and pigs should be controlled at −1%, +4%, and +8.5% or below, respectively. The findings suggest that by 2035, the total carbon emissions of these three animals in the region will reach 696,165 t.

In Gannan’s livestock industry, cattle carbon emissions are the main source of local carbon emissions. In recent years, to reduce livestock carbon emissions and achieve sustainable development in both the livestock economy and ecological protection, the region of Gannan has actively promoted the implementation of grass–livestock balance policies, as well as grassland ecological protection subsidy and incentive policies. By determining livestock numbers based on grass availability, Gannan Prefecture has effectively controlled the stock levels of large livestock. Over the past 16 years, carbon emissions from livestock animals in the region have decreased by 14%. Artificial grass planting for the ecological restoration of degraded grasslands has been conducted, with the comprehensive vegetation coverage of grasslands increasing to 97.14%, playing a significant role in enhancing the region’s carbon sink capacity. In the future, various additional measures can be adopted to further enhance the carbon sink and construct a win–win model for ecological protection and local herders’ livelihoods. Therefore, it is crucial to continue to determine livestock numbers based on grass availability, control the stock levels of large livestock, and moderately increase the numbers of sheep and pigs. The production model of utilizing grass availability to determine livestock numbers and optimizing the quantities of different types of animals is also applicable to animal husbandry production systems in other regions.

The simulations and calculations employed in this paper are solely aimed at achieving a year-by-year decrease in carbon emissions from the livestock production system, with the intention of determining the optimal number of farmed animals. However, other factors such as animal prices have not been considered. To achieve an optimal balance that will minimize carbon emissions from farmed animals in the production system while continuously increasing herders’ incomes, numerous factors, such as the market prices of animals and meat, need to be taken into account. Future research should also explore market-oriented mechanisms, such as grassland carbon sequestration accounting and trading, and green livestock product premiums, to achieve the green and sustainable development of animal husbandry.

6. Conclusions

From 2009 to 2024, the average annual GHG emissions from animal husbandry in the region of Gannan were 774,286 t C-eq, and the overall carbon emissions followed a gradual downward trend. Compared with 2009, the carbon emissions in 2024 decreased by 14%, with an average annual growth rate of −1.0%. Carbon emissions from cattle are the main source of local GHG emissions. To achieve the sustainable development of the animal husbandry economy and protect the ecological environment, the region of Gannan has been issuing and implementing policies for grassland protection since the 1990s, including grassland management methods and grass–livestock balance management methods. The region has also launched subsidies and reward policies for grassland ecological protection. By continuously reducing the breeding of large livestock, controlling the inventory of large livestock, and simultaneously ensuring the ecological restoration of degraded grasslands through artificial grass planting, the region has increased the comprehensive coverage rate of grasslands to 97.14%. At the same time, as the forest area has continued to increase, its carbon absorption has gradually increased, with an annual growth rate of 1.7%, which plays an important role in improving regional carbon sinks. Maintaining a reasonable animal density will help to protect the environment. Future predictions indicate that the rational regulation of the livestock industry structure and the optimization of different animal breeding numbers are also important means to reduce carbon emissions.